[R-sig-eco] Permutation test for spatial autoregressive model

[冯广]

Dear all, 

I would like to ask for some advices or references.

I am interesting in the relationship between two variables. As for their spatial autocorrelation, l would like to use spatial simulatious autoregressive (SAR) model to test this. As far as I know that F-statistic are often compared for OLS results in monte carlo tests. Problem is that, l want to perform a permutation test to determine the significance of the results derived from SAR, while don not know what can be compared between the observed and the expected by null model. 
Log likelihood, pseudo R2 or something else?  

 l am a bit at a loss and thanks in advance for any thoughts. 

Feng Guang
	[[alternative HTML version deleted]]